In wireless communications systems, there are a number of technical limitations. A first limitation is that the frequency spectrum is a scarce resource, which should be efficiently used. For a finite amount of spectrum, there is an upper bound on the number of subscribers that could be simultaneously served. To increase the number of subscribers, multiple access techniques have been introduced in the past.
The most common are: Frequency Division Multiple Access (FDMA), wherein only a small portion of the available spectrum is allocated to a subscriber; Time Division Multiple Access (TDMA), wherein a subscriber is not allowed to transmit continuously, but instead, the subscriber is only allowed to transmit during short non-overlapping periods of time called bursts; and Code Division Multiple Access (CDMA), wherein the total spectrum is allocated to all of the subscribers, who are differentiated by the use of allocated orthogonal codes.
Such and other multiple access techniques are combined in existing wireless systems to maximize the number of subscribers for a finite amount of resources (time, frequency, code, etc.).
Another limitation of wireless communications systems is the finite transmission power that results from overcoming implementation and propagation losses between a transmitter and a receiver. As a result, practical systems have only a finite communication range.
To overcome these two limitations, the cellular concept has been introduced for wireless systems. To cover a large area, the available resources are used for a small coverage area, called a cell, and repeated for other cells. The expected number of subscribers that can be served by a network will increase in proportion to the number of cells in the network.
However, because cells are now quite close together, there is an increased risk of co-channel interference, which will decrease the link quality and commensurately, the number of subscribers.
A number of techniques for combating co-channel interference have been proposed and implemented. These are generally specific to a particular multiple access scheme.
For example, with CDMA, the signals of all of the subscribers within a cell are sent by a base station transmitter in the downlink direction at the same time, so that each specific subscriber may decode its signal and cancel out the intra-cell interference. In the uplink direction, a subscriber's signal is typically scrambled by a long code with good correlation properties, so that the contribution of other subscribers to an individual subscriber's signal will more likely behave as white noise rather than significantly degrade single user detection.
In the case of FDMA systems, the total spectrum is divided into K subsets of frequencies and every cell uses one such subset. Rather than deploying the total spectrum for every cell in the network, a cluster of K cells will be repeated in the network, each being assigned one of the frequency subsets. Clearly, within a cluster, subscribers will not experience any co-channel interference.
For a frequency re-use factor K greater than one, co-channel cells, that is, cells assigned the same frequency subset, will not be adjacent to each other and thus, interference across the network should be minimized. Preferably, the frequency re-use factor is small in order to maximize the number of subscribers, as more frequencies may be allocated within a frequency subset.
Given the recent increase in the number of Base Station System (BSS) features introduced for use by base transceiver stations such as, power control, discontinuous transmission, fractional frequency loading and frequency hopping, an optimal frequency re-use factor may be K=3, with 100% frequency loading.
In any event, to further improve spectrum efficiency of cellular systems, a sectorization concept has been introduced in which an omni-directional antenna, traditionally placed in the centre of a cell, has been replaced by a plurality of N directional antennas, each defining a symmetrical coverage area. Thus, for the same area, the number of cells, and consequently, the number of subscribers within the network, has been increased by a factor of N.
The use of directional or sector antennas has thus further reduced the amount of interference in the network and has resulted in more spectrally efficient networks. A sector is symmetrical and generally wedge-shaped, with N sectors generally extending outward from the traditional centre of a cell. Each sector may now be considered a distinct cell, with its antenna extending from an extremity thereof.
Although, in theory, high spectral efficiency is achievable with large values of N, practical deployment considerations will generally limit this number to a finite set of possibilities. For example, large values of N will cause a significant proportion of the subscribers to languish in continuous handover situations. As well, a cell is generally identified to a subscriber through an identifying code, frequency channel, so that a subscriber may make continuous measurements to identify the best serving cell. With large values of N, a significant portion of the available bandwidth would be allocated to such control channels, without any significant increase in capacity. Consequently, typical values for N are 3, on rare occasions 4, and hardly ever exceeding 6.
When N=3, antennas with a half power beam width of 65° are typically used, because they provide better coverage. For uneven traffic between sectors or for other values of N, multiple antennas may be used with beam widths of 33, 45, 65, 90, 105, etc. For higher sectorization, that is, N>3, a mix of existing antennas will not provide optimal coverage, resulting in either a significant and excessive overlap between beam patterns or else high cusping loss between adjacent beam patterns. In the former case, an excessive number of subscribers will be candidates for handover, while in the latter scenario, coverage holes could result in handover failures.
Furthermore, it appears that the need for higher order sectorization is primarily a local phenomenon, rather than a characteristic across a network, since subscribers are not generally uniformly distributed across a network. As a result, the need for increased subscriber capacity is only apparent for a few scattered sectors in a network that typically encounter large distributions of subscribers.
In such a case, blindly increasing the number of sectors for all of the sites will not result in an efficient capacity to cost ratio since some of the additional transceivers will never be used.
A traditional means of increasing network capacity, known as cell splitting, is to reduce the coverage of existing cell sites and to introduce a new cell site in the newly created coverage holes. Cell splitting is very expensive for an operator, however, since new locations for the tower and equipment for the new site, such as high-rise buildings, have to be located and leased. In many dense urban environments, where increased network capacity would be beneficial, it is no longer possible to find suitable new site locations.
Therefore, alternative means of increasing network capacity are under investigation, such as deploying antennas with optimized beam patterns.